Is the Relationship Between the Credit Default Swap Spread and Distance to Default Dependent on the State of the Economy?

Dennis Dyg Jørgensen Hannesbo

Student thesis: Master thesis

Abstract

We examine the relationship between the credit default swap spread and the Merton distance to default model, which is based on the Merton bond pricing model (1974). We examine the relationship for the estimated default probability and for the estimated distance to default. For both variables we find, in a first difference regression setting, that we must lag them by one period for them to be significant predictors of the CDS spread. The results of our study indicate that distance to default explains slightly more of the variability in the CDS spread than the probability of default. We test if the sensitivity of the CDS spread to the Merton DD model changes when the economy goes into recession relative to normal times. We examine this by including interaction-terms. For the default probability the results indicate that the change in the relationship is insignificant. However, for the distance to default the results show that the first difference of CDS spread becomes less sensitive to changes in the first difference of distance to default lagged one period when the economy is in recession. Furthermore, we examine if the CDS spread can be explained by S&P500 index and the VIX index and if the relationships depend of the state of the economy. The S&P500 index is a significant predictor of the CDS spread, but as for the PD and DtD the first difference of the S&P500 index must be lagged by one period to be significant. The change in relationship between the S&P500 index and the CDS spread from normal times to crisis is insignificant. The conclusion drawn for the S&P500 index is the same for the VIX index in the univariate setting. However, in the multivariate setting the VIX index is insignificant whereas S&P500 remains a significant predictor of the CDS spread.

EducationsMSc in Finance and Investments, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2018
Number of pages92